Antimicrobial Susceptibility and Molecular Mechanisms of Fosfomycin Resistance in Clinical Escherichia coli Isolates in Mainland China

Escherichia coli is one of the most common pathogens in nosocomial and community-acquired infections in humans. Fosfomycin is a broad-spectrum antibiotic which inhibits peptidoglycan synthesis responsible for bacterial cell wall formation. Although low, the exact E. coli susceptibility to fosfomycin as well as the mechanisms of resistance in the population from Mainland China are mostly unknown. 1109 non-duplicate clinical E. coli strains isolated from urine, sputum, blood and pus samples in 20 widely dispersed tertiary hospitals from Mainland China were collected from July 2009 to June 2010, followed by determination of minimum inhibitory concentrations of fosfomycin. Detection of the murA, glpT, uhpT, fosA, fosA 3 and fosC genes was performed in fosfomycin non-susceptible E. coli strains and conjugation experiments were employed to determine the mobility of fosA 3 gene. In this study, 7.8% (86/1109) E. coli strains were fosfomycin non-susceptible. Amino acid substitutions in GlpT and MurA were found in six and four E.coli strains, respectively, while the uhpT gene was absent in eighteen E.coli strains. Twenty-nine isolates carried the transferable plasmid with the fosA 3 gene at high frequencies of around 10−6 to 10−7 per donor cell in broth mating. The majority of isolates were susceptible to fosfomycin, showing that the drug is still viable in clinical applications. Also, the main mechanism of E. coli resistance in Mainland China was found to be due to the presence of the fosA 3 gene.


Introduction
Escherichia coli is one of the most common pathogens in nosocomial and community-acquired infections in humans. Fosfomycin is a broad-spectrum antibacterial agent against Grampositive and Gram-negative bacteria; it inhibits the UDP-N-acetylglucosamine enolpyruvyl transferase (MurA), responsible for the formation of UDP-GlcNAc-enolpyruvate in the biosynthesis of cell-wall peptidoglycans [1,2]. Fosfomycin enters the E. coli cell by using one of two transport systems: the glycerol-3-phosphate transport system (GlpT) or the hexose phosphate transport system (UlpT) [3,4]. The prevalence of fosfomycin resistance in E. coli remains low [5][6][7][8], but several resistance mechanisms to fosfomycin have been reported. Mutations in the glpT or uhpT genes can decrease the uptake of fosfomycin via the transport systems and give rise to fosfomycin resistance [4,9,10]. Resistance against fosfomycin can also be conferred by mutations in the murA gene, resulting in either reduced affinity between MurA and fosfomycin or through the over-expression of MurA [11,12]. In addition, the presence of the plasmid-borne fosA, fosA 3 and fosC 2 genes, which encode fosfomycin-modifying enzymes, can inactivate the drug by catalyzing the covalent addition of glutathione to fosfomycin [13].
However, little is known about the rate of resistance or E. coli mechanisms of resistance to fosfomycin in Mainland China. Thus, here we investigated several fosfomycin resistance mutations and the mobility of the fosA3 gene.

Bacterial strains
A total of 1109 non-duplicate E. coli strains isolated from urine, sputum, blood and pus samples were collected as part of standard patient care from July 2009 to June 2010 in 20 widely dispersed tertiary hospitals. E. coli ATCC 25922 [14] was used for quality control for susceptibility testing and the sodium azide-resistant E. coli J53 [15] strain was used as a recipient in the mating experiment.

Ethics
The samples were collected from hospitals participating in the Ministry of Health National Antimicrobial Resistance Surveillance Net (Mohnarin) study. No ethical approval or informed consent was required due to the retrospective nature of the study. All patient identifiable information was removed from the samples before the authors received them for analysis. The authors were not involved in the treatment and had no direct contact with the patients.

Susceptibility testing
Susceptibility to antimicrobial agents was determined using the agar dilution method according to the Clinical and Laboratory Standards Institute [14]. The following antimicrobial agents were tested: fosfomycin trometamol, piperacillin-tazobactam, cefuroxime, cefotaxime, cefepime, imipenem, amikacin, levofloxacin, and nitrofurantoin. Based on minimum inhibitory concentrations (MIC), E. coli were classified as fosfomycin-susceptible (MIC 64mg/L), fosfomycin-intermediate (MIC = 128mg/L) and fosfomycin-resistant (MIC 256 mg/L). Extendedspectrum beta-lactamase (ESBL) producing isolates were detected according to previously described methods [14]. The reference strain E. coli ATCC 25922 was used as the positive control. The spontaneous fosfomycin resistance rate for E. coli J53 strain determined in our research was <10 −8 .

PCR amplification of genes and sequence analysis
Detection of the murA, glpT, uhpT, fosA, fosA 3 and fosC genes was performed in fosfomycin non-susceptible E. coli strains (intermediate and resistant E. coli strains; MIC 128 mg/L), and the primers used are listed in Table 1 [16][17][18]. Sequence analysis was performed with a Dye primer and a Dye Terminator cycle sequencing kit (Applied Biosystems) and with a 310 gene analyzer (ABI Prism).

Conjugation experiments
The conjugation experiments were carried out to determine the mobility of fosA 3 gene from azide-sensitive isolates as donors to azide-resistant E. coli J53 as the recipient. Overnight cultures of 0.05 ml of the donor and 0.45 mL of the recipient strains were added to 3 mL of fresh Mueller Hinton (MH) broth (Oxoid Ltd., Basingstoke, Hampshire, United Kingdom), then incubated and gently stirred at 37°C for 12 hours. 0.1 mL of the mixtures were plated on MH agar (Oxoid Ltd., Basingstoke, Hampshire, United Kingdom) containing glucose-6-phosphate (25 mg/L), sodium azide (100 mg/L) and fosfomycin (Zambon Group, Milan, Italy) (32 mg/L) and incubated for 48 hours for selection. The conjugative transfer frequency was calculated as the ratio of the number of conjugants to the number of donors.

Statistical analyses
Statistical tests were performed using Social Sciences software for Windows Version 14.0 (SPSS, Inc., Chicago, IL, USA). Enumeration data were expressed as percentage values. The differences in susceptibility between the groups were compared using the Chi-square test or Fisher's exact test. The differences between the groups were considered significant if the p-values were smaller than 0.05 (two-sided test). The Bonferroni method was used to adjust the significant levels (0.05/3 = 0.0167) in multiple comparisons between any two levels of the susceptibility outcome.

Results
1023/1109 (92.2%) of the E. coli isolates were susceptible to fosfomycin. The susceptibility rates of the isolates from urine, sputum, blood and pus samples were 95.0%, 87.1%, 93.3% and 93.3%, respectively. MIC 50 of the strains from different specimen types were all 0.25 mg/L, whilst the MIC 90 of the strains from urine, sputum, blood and pus samples were 4, 128, 16 and 32 mg/L, respectively ( The antimicrobial resistance rates stratified by fosfomycin susceptibility categories are summarized in Table 3. Fosfomycin-resistant isolates were significantly more likely to be ESBL positive than were fosfomycin-susceptible isolates (P <0.001). In addition, the resistance rates for piperacillin-tazobactam, cefuroxime, cefotaxime, amikacin and nitrofurantoin were significantly higher in fosfomycin-resistant isolates than fosfomycin-susceptible isolates (P < 0.0167 for all comparisons).
In 86 fosfomycin non-susceptible E. coli strains, amino-acid substitutions Val389Ile, Asp390Ala, Gln59Lys and Glu139Lys were found in MurA of E. coli strains J621, E261, E374 and R162, respectively. In addition, amino-acid substitutions Ile4Val and Gly84Asp were found in GlpT of E. coli strains H039 and R538, respectively. In strain Q446, the GlpT protein was truncated due to deletion of nucleotide 401A. Amino acid substitutions Thr144Pro and Pro173Ser were found in the GlpT of strains R046 and P305. Amino acid substitutions Ala12-Val and Gln437Cys were found in GlpT of strains T229 and T297. The uhpT gene was absent from 18 of the test strains. 69 E. coli isolates were positive for fosA 3 , yet no fosC 2 or fosA genes were detected among these isolates ( Table 4). The fosA 3 gene was able to transfer at frequencies varying from 1.1×10 −7 to 9.9×10 −6 between the donor and recipient in 29 isolates tested.

Discussion
Fosfomycin has been introduced in clinical practice for about 30 years. However, this antibacterial agent is still not commonly used in China, and the fosfomycin resistance rate in clinical E. coli isolates remains very low [18]. In this study, only 7.8% (86/1109) of E. coli isolates were fosfomycin non-susceptible with a high ESBL-positive rate and levofloxacin-resistant rate. Among the fosfomycin-resistant isolates, the ESBL-positive rates were higher than in fosfomycin-susceptible isolates. Fosfomycin inactivates the MurA enzyme by binding to the active site of its Cys-115 residue [19,20]. The amino acid Asp369Asn and Leu370Ile substitutions have been reported in fosfomycin-resistant E. coli isolates MSC17327 and MSC17323. Inspection of the crystal structure of E. coli MurA complexed with fosfomycin does not suggest an obvious role for Asp-369 and Leu-370 in the protein-inhibitor interaction [16]. In our study, the MurA substitutions of Val389Ile, Asp390Ala, Gln59Lys and Glu139Lys were found in four E. coli strains. However, further investigations are needed to find out whether the substitutions contribute to fosfomycin resistance. Fosfomycin is transported into cells via two pathways: the glycerol-3-phosphate or the hexose phosphate transport systems. Several studies have reported E. coli fosfomycin resistance due to the defects in GlpT or UhpT [9,13,16]. In this study, mutations in the glpT gene were found in six E. coli strains, all of which resulted in amino acid substitutions in GlpT. In addition, single nucleotide deletion in the glpT gene, which would lead to a truncation of the GlpT sequence, was also detected.
Alteration in the chemical structure of fosfomycin by FosA 3 , a protein encoded by the fosA 3 gene, was previously reported in three E. coli strains in Japan in 2010 for the first time [13]. In our study, over 80% of fosfomycin non-susceptible E. coli strains (69/86) harbored the fosA 3 gene, which indicated that it is the main mechanism responsible for fosfomycin resistance in Mainland China. The fosA 3 gene identified in 42% isolates (29/69) can transfer between different E.coli strains. Previous research has suggested that the fosA 3 gene is encoded on a conjugated plasmid [21,22]. As the mobility of this gene may accelerate the dissemination of fosfomycin resistance around the world, future research is warranted to confirm this for E. coli.